Research Grants 1998-1999

Bioartificial Human Liver Lobule

Jeffrey M. Macdonald, PhD
University of North Carolina, Chapel Hill, North Carolina

This project combines the emerging field of "tissue engineering" with the non-invasive technology of nuclear magnetic resonance (NMR) to create a toxicodynamic bioassay for determining mechanism(s) of toxic action in normal human liver tissue. Macdonald and colleagues have constructed a hollow fiber coaxial bioreactor with the same dimensions as a liver lobule, which is compatible with NMR spectroscopy and magnetic resonance imaging (MRI). Ultimately, they plan to culture fluorescent activated cell-sorted human hepatic progenitor cells in this coaxial bioreactor and train these liver cells to form bile ductules which will attach to the bioreactors bile ports. However, initially the team will culture a human hepatoma cell line, HepG2, in order to optimize bioreactor flow dynamics, culture conditions, and multinuclear NMR spectroscopy. NMR spectroscopy and MRI are non-invasive techniques that permit one to "tune -in" to radio frequencies emitted by various nuclei. Multinuclear NMR spectroscopy determines toxicodynamics by monitoring changes in metabolite intermediates. Macdonald will use ¹³C NMR to follow glycolosis, gluconeogenesis, TCA cycle, glutathione, pentose-phosphate pathway, or any other changes in metabolic pathways that arises due to the effects of toxicant. This is possible because NMR does not selectively analyze for specific compounds like mass spectroscopy or chromatography, and it is non-invasive permitting "real-time", in vivo analysis. ³¹P NMR will be used to monitor bioenergetics (i.e., changes in ATP levels), ¹⁵N NMR will be used to monitor the urea cycle, and ¹⁹F NMR will be used to monitor P-450 activity and signal transduction. In studies with HepG2, ¹H MRI microscopy will be used to identify diffusion and perfusion in the bioreactor at various flow rates. Using a perfluorinated compound, ¹⁹F MRI microscopy will be used for determining oxygen dosimetry across the cell mass at various flow rates. The empirical data will be determined. Oxygen concentration profiles as a function of flow rate will be generated and serve as a guide for the expensive and precious human hepatic progenitor cell culture.

The bioartificial human liver lobule model is part of a larger toxicity testing scheme that incorporates whole animal, cells, and macromolecular NMR models to determine mechanism(s) of toxic action. The animal (Macdonald et al., 1998) and the macromolecular (Macdonald et al., 1998) models are established, but the cell model (i.e., bioartificial human liver lobule) still needs characterization. The toxicity testing scheme relies on a neural network to relate spectral changes induced by administration of a toxicant to the human hepatocytes to potential targets. Then the macromolecular model determines the interaction of toxicant with the target, identifying reactive functional group(s) on the macromolecule. Based on the targets of the toxicant, probable target organs are determined, and the animal NMR model is used to verify the mechanism of toxic action. This overall scheme reduces the use of animals in research because the bulk of studies to determine the mechanism(s) of toxic action use cell cultures of primary human hepatocytes. Due to the non-invasive aspect of NMR spectroscopy, many animals are not required per time point; rather a single individual can be followed through the entire experimental timecourse. In theory, this reduces the number of animals in a standard toxicology study by a factor of:

(the number of timepoints per study x number of animals per timepoint) - (1 x number of animals per timepoint).